| PID controller has simple structure and good robustness,there is extensive application in many fields at present.But with the development of science and technology,the controlled system has become more and more complicated,so the conventional PID controller can't get good performance for the time-varying objects and non-linear systems.And the neural network has the ability of adaptive and self-learning,the traditional PID controller is transformed by the neural networks,there is a better control result to the complicated system in industrial control,and improving the instability of the result because of the structure and the parameter of the system change.In this paper,the neural network adaptive PID controller for the simulation study,mainly in the following aspects:(1) Detailed description of the neural network theory and learning algorithm of neural network,leads to the single neuron adaptive PID controller and the neural network adaptive PID controller based on Back Propagation algorithm.(2) Carried on simulation to the single neuron adaptive PID controller and the neural network adaptive PID controller based on Back Propagation algorithm,the simulation results show that the neural network adaptive PID controller not only has the advantage of traditional PID but also has the features of self-adaptive.It can regulate the system in time when the object changes or something interfere the system ensure that system is carried out smoothly.(3) Demonstration the superiority of the neural network adaptive PID controller,used the single neuron adaptive PID controller in the DC motor system.The simulation analysis,the DC motor system to follow performance,anti-around performance and robust performance has been improved and has reached the expected results. |
